Understanding of the mortality transition has been limited by what has been called a false precision of cause-of-death data in historical and data-deficient populations. These problems have, in turn, placed limits on the analysis of cause-specific transition hypotheses such as the role of specific medical interventions in mortality transition and in the use of automated nosology in such settings. This proposal addresses problems with the historical precision of cause-of-death data through an integrated archival social history of death reporting and analyses of changing historical cause-of-death nomenclature, social biases in the reporting of deaths, and an analysis of probabilistic cause-of-death classification, resulting in the estimation of robust cause-specific mortality trends for major causes of death and selected causes of importance to understanding the later nineteenth-century mortality plateau and eventual mortality transition in America. This research uses the case of the American epidemiological transition for a formal analysis of literal causes of death in the Connecticut River Valley mill towns of Holyoke and Northampton, Massachusetts from 1850 to 1912, during the critical onset of the North American epidemiological transition, the sanitation movement and germ theory, and the first modern mortality reporting system in the United States. Results of the mortality analysis are compared to those of historical and automated conventional nosologies. [unreadable] [unreadable] [unreadable]